1

1

Legacies of the Tokugawa Regime

• Establishment of the Tokugawa Bakufu– Tokugawa Ieyasu (1542-1616)

• Important Policies– Position of Emperor– Banning of Foreign Trade– Bakuhan system– The Rise of Cities

2

Japan Prior to 1600

• Japan had interacted heavily with China– Chinese Characters – Political Ideology was Confucian– Religion was a combination of native animism (Shinto)

and Buddhism, which originated in India and entered via China

• Between 1500 and 1600 Japan was immersed in a long sequence of bloody civil wars– In 1600, Tokugawa Ieyasu emerged as victorious at the

battle of Sekigahara– In 1603 the emperor named Tokugawa Ieyasu shogun

2

3

Political Issues

• Tokugawa government feared local rebellion– Prohibited local population from using guns– Wanted to make sure people would stay put and

produce

• Divided the Daimyo into two types (approx)– 150 fudai daimyo - hereditary retainers - those that had

allied themselves with Tokugawa– 100 tozama daimyo - outside lords- those that had

opposed– Tozama han were placed between fudai han in outlying

areas

4

Tokugawa Policy• Shogun ruled in the name of the emperor

– Emperor continued to remain as a figure head• Banned all Foreign Contacts from 1639 to 1856

– No Trade except with Dutch and Chinese• Established the Bakuhan system

– Bakufu had control over the daimyo (feudal lords) • Daimyo had control over their han (feudal domains)

– Daimyo were basically administrators of regions that produced more than 10,000 koku (Japanese bushels)

– Han were relatively small• In 1598 Japan produced about 18.5 million koku• Shogun controlled about 2 million koku• Only one had more than 1 million Koku• Only 22 had more than 200,000 koku

3

5

Social Structure• Adopted a four class social structure based on

Confucianism– Samurai, farmers, artisans and merchants– System similar to Physiocrats in France: “All value comes out of

the ground”• Peasants constituted 80% of the population and were

heavily taxed to support Daimyo– Peasants were not permitted to leave farms but often did– Small Scale farms– Little capital

• Little use of machinery until end of nineteenth century• Merchants were just “movers of goods” and therefore

didn’t have real value

6

Development of Transportation Networks

• Established Sankin Kotai (alternate attendance) system– Required daimyo to spend every other year in capital Edo (Tokyo)

and leave family behind– Caused roads, communication networks, towns to be built along

the Tokaido road.• Five major roads left Edo• Dutch were very impressed, indicating relative development of Japan• Postal system became quite advanced• Greatly enhanced the wealth of Edo• Edo was probably Largest city in world by 18th century with a

population of over 1 million• Population of Kyoto was over 500,000• Osaka had a population of 400,000

– Relatively clean and well run

4

7

8

Urban Policy• Bakufu issued decree requiring each daimyo to build a

castle in his han– Samurai had to be kept in the castle town– Mechanism of monitoring weaponry– Prevented samurai from marauding and stealing from peasants– Broke connection between Samurai and land– Broke connection between Samurai and warrior past– Made Samurai dependant on Daimyo for stipends b/c couldn't

collect them directly.

• Contrast with Europe– In Europe the aristocracy ruled from countryside as opposed to the

urban based system in Japan– More urbanization than Europe

5

9

Legacy of the system today• It is often argued that reation of “castle towns” in

Tokugawa period created agglomerations of people

– These cities were more productive and created inward migration

– Industries changed but the centers of production did not

– Had the capital not moved to Edo, Tokyo would have remained a backwater.

10

6

11

12

7

13

14

8

15

Theory 1: Random Growth • Random Growth

– “History is one damn fact after another” – Henry Ford– Big cities form when the random events of history conspire to

make them a location where many people end up• Immigrants arrived in New York and tended to stay here• People killed in catastrophic events result in permanent drops in

population of the same magnitude.

• Not a Lot of Economics Underlying This– Oddly enough, this theory is successful at explaining the rank-size

rule (Zipf’s law)– The rank-size rule is the fact that in virtually all countries the size

of the nth largest city is 1/n times the size of the largest– Stunningly robust result in most countries.

16

Theory 2: Locational Fundamentals• Locational fundamentals posits that there are

unchanging characteristics of locations that determine the size of cities– For example, proximity to rivers, coasts, harbors,

deserts, mountaintop, rainfall, flatland, latitude, etc

– Each location is the sum of a long sequence of randomly distributed spatial qualities

– Difference with RG is response to temporary shocks:• Random growth: shocks are permanent• Locational fundamentals temp shocks have no long run impact

9

17

Theory 3: Increasing Returns to Scale• Productivity of firms is enhanced by surrounding

economic activity– Achievement of Minimum Efficient Scale (MES)

• Not enough local demand to achieve MES, so firms are less productive

– Informational Exchange: • The more people one interacts with, the smarter one gets (and

hopefully) the smarter they get! • But it is harder to interact with people who are farther away

– Cost Linkage• The more varieties of inputs a firm has the better the output• But imported inputs are more expensive than local inputs

– Cities arise because we are more productive when more concentrated

18

How to Measure Regional Dispersion?• Concentration Ratio: Share of Population in

Largest Regions

• Relative Variance– Ratio of Variance of Log Population Density in Year t

Relative to 1998

• Zipf Coefficient– Coefficient in Regression of Log of the Rank of a

Region on the Log of Its Population Density• Note if region n 1/n times the size of region 1 then

– Pop(n) = 1/n * Pop(1) or ( ) ( ) ( )ln Pop ln Pop 1 lnn n= −

10

19

Size Distribution of Feudal Domain Size in Japan (1597) Top 50 Daimyo

Zipf's Law for Daimyo

y = -0.9326x + 14.224R2 = 0.9801

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

11 11.5 12 12.5 13 13.5 14 14.5 15

ln(Rice Output of Han)

ln(R

ank

of H

an)

20

Zipf's Law and US Cities

y = -1.0012x + 10.547R2 = 0.9869

0

1

2

3

4

5

5 6 7 8 9 10

ln(size)

ln(r

ank)

11

21

Zipf’s Law and Japanese Regions

Log

Ran

k of

Pop

ulat

ion

Den

sity

Log of Population Density

Log Rank of Population Density Fitted values

4 6 8

0

1

2

3

4

Slope = -0.96, R2 = 0.98

22

Evaluating Theories: Historical Regional Population Data

• Data for the Years 725 –1150 Is Based on Censuses Done for Tax Purposes– Japanese Tax System Was Lump Sum and Therefore Required

Detailed Censuses and Tax Districts Based on Population

• Data for 1600 Is Augmented With Major Land Survey

• From 1721 to 1998 We Do Not Need Tax Data As Actual Population Data Exists

12

23

What has happened to regional density?Look at Variation across 36 Regions

Jomon 125 0.39 2.46 -0.809 0.53 0.31Yayoi 595 0.23 0.93 -1.028 0.67 0.50725 4511 0.20 0.72 -1.207 0.60 0.71800 5506 0.18 0.75 -1.184 0.57 0.68900 7442 0.29 0.68 -1.230 0.48 0.651150 6836 0.20 0.66 -1.169 0.53 0.731600 12266 0.30 0.64 -1.192 0.76 0.831721 31290 0.21 0.43 -1.582 0.85 0.841798 30531 0.21 0.37 -1.697 0.83 0.811872 33748 0.18 0.30 -1.877 0.76 0.781920 53032 0.25 0.43 -1.476 0.94 0.931998 119486 0.41 1.00 -0.963 1.00 1.00

Raw Correlation with 1998

Rank Correlation with 1998Year

Population in

Thousands

5 Largest Region Share

Relative Var of log Pop Den

Zipf Coefficient

24

Persistence of Density

rank

of (

d199

8)

rank of (d1600) 0 10 20 30 40

0

10

20

30

40

13

25

What did we learn?• High degrees of variation in population density has

always been a feature of the world– Zipf’s law has always held, i.e. nth largest region is 1/n times the

size of the largest region• Suggests the importance of either Random Growth model or

Locational Fundamentals Model– High degree of persistence in the size of regions over time

• Suggests that Locational Fundamentals are critical to understanding region size

– There has been an increase in concentration in last century which coincides with industrialization• Increasing returns may play a role in the size of cities